To get away from the HDR photography for a little while I’ve added a new page where I’ll be posting high speed videos of anything interesting I shoot. One of the great advantages of still being in college is the opportunity to use all the cool equipment lying around - in my case a high speed camera and a gas powered gun! I’ve posted one video so far - you can see it here. With any luck I’ll be able to shoot a few more over the next couple of weeks; I’ll post them soon….
I’ve just noticed that a lot of images have vanished from this site - It’s rather ticked me off as I don’t have any backups with me at the moment. All images should have had ALT tags with them so you hopefully have some idea what the images were. Sorry for the inconvenience - I’ll restore all images within 24 48 hours. (Starting…. NOW!)
Breaking News: As far as I can tell all images have been restored. HURRAY! Think I’ll go make some backups…
One of the necessities of creating pseudo-High Dynamic Range photographs is having multiple exposures of the same shot. This gives us the full tonal range of a natural scene needed to blend a HDR together. Blending is followed by tone mapping, which maps the colours contained in the dynamic range of the HDR to a rough set of similar colours that can then be displayed on monitors or any equipment that doesn’t have the full dynamic range to view the HDR in its true form properly. The question that follows then is how many different exposures do you need to take? And across how wide a range? ±1EV? ±2EV? Some people might go as far as ±4EV to be sure they capture every single shadow and light. Now while combining the metered image with 2, 4 or 6 exposures will give us a higher dynamic range image, I wanted to see how well the HDR process works using just a metered image (i.e. at 0EV) along with just two differently exposed duplicates; taken at ±1EV, ±2EV and ±4EV. For this example I’ve used Photomatix to create the mock HDR.
Here is each exposure as taken from the single RAW image in Photoshop (I’ve said it before - I never have a tripod with me so taking 3 or more RAW shots in near perfect alignment is impossible for me!). The exposures range from -4EV to +4EV with the metered image sitting in between them.
1. The Original (Metered) Photograph
First things first. Below is the original shot saved as a jpg from the RAW. As you can see the shot is generally a little over-exposed (the sky very) and most features are looking a little bland. A bit a post-processing is necessary me thinks.
2. Photomatix Using Just The RAW File
So the first step in this experiment is to run the single RAW image, no multiple exposures, through Photomatix and try to keep everything looking at least half realistic: The Photomatix results using just the RAW image came out reasonably well below: the colours in the building and the statue are stronger and give the picture more life, without looking over-saturated. The sky however looks washed out, almost as if it was part of a water painting. So Photomatixing a single exposure doesn’t seem to give great results, at least for this example - on to using our metered shot and ±1 exposed shots.
3. Photomatix Using RAW And Just ±1EV Exposures
Using the ±1EV exposures along with the RAW file has definitely improved the situation: there is a little more tonal range on the statue and across the building compared to the Photomaxifed RAW, and the sky is showing much improvement. I posted this photo over at The Photo Forum for some opinions. Now lets try ±2EV.
4. Photomatix Using RAW And Just ±2EV Exposures
The results of using ±2EV Exposures are similar to that of the ±1EV photo. We can tell though that the colours in the sky are slightly more contrasted / saturated. One improvement I think is the shadowing on the statue’s leg, which has been reduced a little due to the greater exposure range.
5. Photomatix Using RAW And Just ±3EV Exposures
OK now things are beginning to go downhill! Although the rest of the shot is holding up well, the sky has been washed out due to using too wide a range of exposures; remember the ±1EV or ±2EV have not been used here so we are missing all the information contained in those exposures - the gap between the metered shot at 0EV and exposures at ±3EV has grown too great. We need the information in these exposures to accurately capture the proper tones. Let see though what happens if we increase the exposure range even more!
6. Photomatix Using RAW And Just ±4EV Exposures
Well this is just a progression from the photo before - the sky is washed out even more. I’m surprised though that the rest of the image is more or less unchanged, I guess it has something to do with the original composition of colour and contrasts but I’m just guessing.
When taking multiple exposures for a HDR image and you only wish to take one on either side of you metered shot, there’s no real need to choose huge ±EV number: 1 or 2 seems to suffice: my S9600 offers auto braketing only up to ±1EV which I always thought was too narrow, until now! Go beyond 1 or 2EV and you risk saturating your photographs with cartoon like textures. If you like taking 5 or 7 shots for merging however, you can disregard this post!
Related Articles:
Tutorial To Create A High Dynamic Range Photo (HDR) From A Single RAW Image.
Creating Ultra Vibrant Tone-Mapped Surreal HDR Photos With Photomatix Pro
While I would prefer to provide a full RSS feed for this site, for the immediate future I’m sorry to say I can not. Wordpress is just great - each post one makes is automatically added to ones feed at http://yoursite.com/feed. So in my case it’s located at http://projectvisual.net/feed. Unfortunately when I set my feed to full posts, Wordpress also includes my Google Adsense ads that are at the beginning of each post. Now I can’t imagine anything more annoying than subscribing to a feed only to find adverts in them - I’d drop that feed without a seconds hesitation!
At the moment I’m unsure how to easily edit this feed, or if there’s a plugin to disable the Adsense code from being included. I could just make the feed manually but availing of the automatic feed that Wordpress provides is just so fluidic I’m hesitant to give it up! So for the moment I’ll just be providing a summary feed (there’s no Adsense code in these) and as soon as I figure out how to disable the code from the full feed they’ll be back up again.
A 28mm wide angle lens has a viewing angle of around 76° degrees, a circular fisheye lens will give you a 180° viewing angle, and a wacky distorted image, but as far as I know there’s no such thing as a 360° lens (yet). This limitation makes it a little difficult in creating 360° panoramic images of your favorite horizons. Fortunately if you take stand in one spot and rotate incrementally while taking multiple images, each one overlapping the next by about a quarter, they can be imported into Photoshop (CS3 anyway, I’m not so sure about earlier versions) and merged together quite seamlessly.
The procedure is simple but there are a few options to choose from. First of all we have to choose our images. To illustrate I’m going to use some photos I took while on the top of Benbulben (a small mountain in Ireland). The final image itself is far from spectacular: it’s mostly just grass, but that’s not the point here - I want to show you how overlapping images (of any sort) can be combined. Here are the images I’ve chosen to merge:
With Photoshop open go to File -> Automate -> Photomerge. We are now presented with this box:
Immediately we can see on the left five options for merging our images and options for selecting the images themselves. For most people, myself included, choosing ‘Auto’ from the left-hand-side options is usually the safest way to go - Photoshop does a great job in stitching images together and most of the time the process is seamless. For completeness sake however I tried combining my images using each process to see what results I could get.
1. Auto
As expected Auto merging has done it’s job well: from the image below we can see, or rather can’t see, any sign of stitching between each photo. What Photoshop has done here is place all the image files onto their own layer, on one large canvas (see right) and searched for similarities between each image. Depending on how much similarity Photoshop finds between each one, it arranges them accordingly and applies a mask layer to each layer. Finally Photoshop removes any part of the photo that isn’t useful, usually any sections near the interface of two fo the images, as well as the rest of the canvas for that layer. This can take a while for Photoshop to do - for this example it took about 10 minutes or so. Choosing the auto option has the advantage of using a ‘click and forget’ approach: it gives very good results but you can’t fine tune the image positioning if you feel you need to. Personally I think the panoramic shot below would be far better if I could centre the view on the right had side, leaving the grassy sections on the outer edges of the shot. I’ve left the image uncropped so you can see how much you can typically expect to lose from each photo when you actually do the cropping.
2. Perspective
Tyring to photomerge the images using the Perspective option didn’t fare so well - I got this message half ways through the merging process:
Some images could not be automatically aligned.
Choosing perspective is useful if there is a good sense of distance in your shot. Using the Interactive Layout option later on I was able to apply perspective to some of the images (albeit with another error) that you can see below.
3. Cylindrical
If you wish to create a panoramic image in which you want to exaggerate curvature, for example the horizon of the earth shot from a plane or a mountain top, then you should try using Cylindrical Photomerging. I don’t think it applies to the photos here but I’ve used it quickly anyway for completeness sake.
4. Reposition Only
The reposition only option allows you to move each image once Photoshop has finished merging them, so if for example you feel one could be positioned a little bit better you are free to do so, but Photoshop won’t re-blend afterwards which means you’ll have to alter the mask layer yourself if needs be. Sounds like a lot of hassle for very little gain in my opinion. You can see in the thumbnail below the gaps between each image - they won’t show up after you save
5. Interactive Layout
The interactive layout option gives you full control in positioning each image and applying perspective. It will also try to blend each image again after repositioning, a big advantage over the Reposition Option.
You can easily see from the image below that even after I move each photo completely out of place Photoshop will do its best and try to blend them together anyway - something that won’t happen using the Reposition Only option.
I tried applying perspective here and while I still got an error…
Photomerge was not able to automatically correct the perspective for all of the images.
…photoshop allowed me to remove a couple of image and I was left with whats below. Unfortunately Photoshop has completely blown the long shot image in an attempt to add the perspective. I suspect this is because there is such a sharp change in distance from one shot to the next.
So if you have multiple photos then that you wish to merge into one long panoramic scene then you can try some of the option here that Photoshop has available. You can see the final shot here; like I said earlier it’s mostly just grass! I’ve made a slightly more interesting one from the top of Croagh Patrick that you can see here also!
Related Articles
No one likes watching movies on a small screen. While my 13inch macbook provides for some crisp resolutions it doesn’t exactly make for a cinematic experience when watching my favourite DVD. To this end I connected my macbook up to my 32inch LCD HDTV but alas I couldn’t at first find a way to display the movie on the television in fullscreen mode - I could only stretch the movie to fit the screen. As always though with a bit of perserverence the answer will present itself to you, and like most things there’s actually no difficulty involved; it’s purely a matter of knowing how…
Physically connecting a macbook to a HDTV (or monitor) is very easy: the macbook has a mini-DVI port, and HDTVs and computer monitors will have either a VGA, or HDMI port, or both. Mini-DVI is a rather inconvienient kind of port as seems to be only used by Apple so is a tad uncommon. Apple sell a way too short mini-DVI to VGA (female) cable; but a male-male VGA cable will complete the connection to you TV / monitor. The macbook automatically detects the second screen and will do whatever internal configurations are necessary to get the display working. It’s then just left up to you to set the resolution within System Preferences -> Displays. Just be sure that ‘Mirror Displays’ in Display -> Arrangement is unclicked (actually you can simply mirror the displays and run the movie in fullscreen on the macbook screen but it’s likely that the resolutions on both screens are not the same so you’ll be left with a border around the movie. Also you may wish to use one screen while playing the movie in the other so it’s best to unclick the ‘Mirror Displays’ option).
Once the second display is set up, open any movie in Quicktime and then go to View -> Present Movie in the menu toolbar. Here you can select which screen you want the movie to play in and at what size you want the movie to ‘Present at’ (probably Full Screen). Hit play and your move should zoom over to your second screen and begin playing - WALA!
Now this will only work while Quicktime is the active window. As soon as you change to another application, say a web browser, Quicktime will revert back to its original window. To get around this go into the quicktime preferences and select the Full Screen tab. At the bottom is the option to ‘Remain in full screen when player is inactive’. Turn this on and then you’ll be free to do anything you want on your macbok while your DVDs play fullscreen in all their glory!
Hello everyone! I’ve just transferred by website from rapidweaver to Wordpress, since I am hosted by Dreamhost and they offer 1 click installs of all the necessary software. Transferring over will also let me upload new posts from any computer, rather than being confined to using my oWn rapidweaver file on my laptop!
In the meantime why not check out the most recent posts to the right? Maybe what you’re looking for is there!
One thing that most people have to do quite a bit if they have a website with a lot of photographs on it is resizing. It’s great to keep high resolution, high Megabyte originals on your hard-drive, but when displaying them on a website, one must account for their monthly bandwidth allowance from whatever provider they are affiliated to. Even more importantly, they have to be aware that visitors will not hang around long waiting for large files to download. So it’s good idea to know that when it comes to minimizing megabytes, some programs out there are better at resizing an image than other. A lot better. You have a lot of flexibility when dealing with thumbnails also because in addition to them being low resolution, they don’t have to be saved at the highest quality either.
For this little comparison a RAW image (saved size is 18.3MB) with a resolution 3488×2616pixels was converted to a jpg in Photoshop without any other post-processing. The image quality was set to level 10 and saved, giving a new saved size of 2.3MB. I usually give thumbs of pictures an resolution of 400×300pixels so the visitor can still see the image clearly without having to wait long for it to load. In Photoshop, resizing this jpg down to 400×300pixels using the Bicubic Sharper algorithm and saving it at level 4 (low). The resulting file still comes out at 116kb! Far too big for a thumbnail image I think! The image can be seen here.
Using the exact same procedure in GIMP the image came out at 20kb, which is much more manageable. The GIMP output can be seen here. Some of the vibrancy in colour has been lost in the GIMP version but I would argue that this loss in detail is more than compensated for by the 500%+ reduction in file-size. Remember, thumbnails are only supposed to be links to larger, better quality images, they only serve to give the visitor a sneak preview of the full picture. Bear in mind that this was merely a rough and ready comparison between the two programs; no doubt there are numerous settings I could have toyed with for hours to give a different result - I simply went with the default settings for now as they both give reasonable results and you shouldn’t dwell too much on thumbnail quality anyway.
I’ve been experimenting with HDR post-processing for a little while now and I’ve already put up a few images that you can see at HDR Photography. I use Photomatix Pro to apply the tone mapping and while it does add some vibrant colour to what are often initially bland images, they still are not the super surreal, in your face, completely over the top photos that I thought I could achieve. A good example of what I’m talking about can be found at PhotoFencePhotography.com. Fortunately I realized that there’s a really easy way to this - it’s actually so simple that I’m amazed I didn’t realize it before. OK so if you’ve been to other parts of this site you would have come across the post Tutorial to create a High Dynamic Range Photo (HDR) from a single RAW image. Read it? OK lets move on then. Lightroom isn’t essential to create the different exposures; Photoshop is more than able to do this job for a RAW image. For this example I’m going to use a photo I took in Vienna; it’s of the Donner Fountain (Donner-Brunnen) found near St Stephen Cathedral. It’s the most famous fountain in Vienna and gives a great surreal effect once it’s been Photomatixified. You of course can use any image in your own library if you wish. Here’s the original picture….
I’ve uploaded it as jpg here but you should take your photos in whatever RAW format you camera has as you will have much more information within the image to play around with. So, like before create a number of differently exposed photos, I’ve taken five here (+2, +1, 0, -1, -2). This is done to extract the wide range of lights and darks that are available to you from a RAW image, since a RAW file contains all the information that the camera sensor recorded while the photo was taken. This is typically much more than if we were to take our pictures as jpgs, which are compressed.
I choose to make five differently exposed images here; you than do more or less if you prefer. I wouldn’t expect the final result to be significantly altered. OK, once these files are created, we open up Photomatix, and in the menu go to HDR -> Generate and select all our differently exposed images. Leave the settings at their defaults (but change the exposure values that Photomatix suggests if need be) and hit OK to generate the HDR. When that is done, in the Photomatix menu, select HDR -> Tone Mapping and we should get a more vibrant version of our original like below…
Now this looks more true to life - this is more like what I actually saw in Vienna - the wider range of tones, the shadows and the general warmth in the picture is much more appealing. However if we want to forgo realism for the moment, rather than closing the HDR image we just created we can very simply go back to the Photomatix menu and select HDR -> Tone Mapping again! Photomatix will tone map our already tone mapped picture and give us something like this…
Completely unrealistic but very cool none the less. The final result will depend a lot on the actual post-processing that we do in Photomatix (such as color saturation, light smoothing, microcontrast and micro-smoothing). As a quick guide, generally after tone mapping the image for the first time, keep light smoothing high, to reduce halos around objects, and colour saturation and strength near maximum to bring out the vibrancy. For the second tone map light smoothing can be reduced to bring out the halo effect, which gives the picture alot of its surrealism, but the color saturation may have to be dropped. As a final experiment, I tone mapped the image a third time but the only way really to make is look half decent was to drop the color saturation to zero! It looks rather bland here against the grey backdrop but the higher quality version looks pretty good I think!
Tutorial to create a High Dynamic Range Photo (HDR) from a single RAW image.
Vienna
Subscribe to the RSS Feed